I. Methods of taking and examining samples of milk for Str. agalactiae are described.

Where delay in examination is unavoidable Edwards's (1) medium is recommended. This selective medium, however, does not permit the growth of some types of organisms which may cause a subclinical mastitis. It is therefore recommended that, where possible, samples be examined immediately after taking and a non-selective medium used. In this way a measure of the udder count is obtained which is of great value. Where samples can be examined within a few hours plain blood agar is preferred. If blood is unobtainable ordinary milk agar may be used. Non-blood media have certain advantages over blood media, which compensate for the loss of ability to detect haemolysis when it occurs.

A simple confirmatory test for Str. agalactiae is described. Confirmation of type is regarded as essential in all but the obvious cases.

II. The results of a number of simple tests for mastitis have been compared with those from examination in blood agar. These all have roughly the same order of efficiency in that they detect about half the positive cases and give 10–30 % false positives. It is suggested, therefore, that with two exceptions (the milk agar count and the rennet test, which may be used for special purposes) they are so unreliable as not to be worth doing in any laboratory examination.

The brom-cresol-purple paper, strip cup and induration tests, which are very simple and can be carried out by the milker, are worth doing in order to get a general idea of the incidence of mastitis in the herd. Their limitations should be realized, however, for no simple test, or combination of simple tests, is really satisfactory, especially if the owner wishes to secure eradication of the disease from the herd. Our results thus confirm those of other workers. III. Mastitis has a marked effect on the “udder count” as given by total colony count on milk agar. No samples from mastitis-free cows were found to have a count of over 2000 per ml. and very few over 500. About 50% of infected cows yielded milk with a count of over 1000.

The milk agar count is perhaps the best of the indirect tests as although it detects no more than other tests (about 50 %) it gives very few false positives. A count of under 100 per ml. may be taken to indicate freedom from infection and a count of over 1000 to indicate mastitis.

IV. The methylene blue reduction test is not capable of detecting mastitis. Reduction time appears to be more closely related to the cell content of infected milks than to the total count. Storage of the samples for 16 hr., either at 4 or 15·5° C, did not materially improve the efficiency of the test.

V. The possible causes of the unsuitability of mastitis milk for cheesemaking are classified and discussed. The brom-cresol-purple-rennet test is described and recommended as a general test for mastitis and suitability of the milk for cheese-making. The effect of mastitis on the chemical composition and enzyme content of the milk is discussed. Mastitis adversely affects the “body” of cheese and is an important factor in the fault known as “red spot”. Many samples of mastitis milk give slow growth of starter, and these are usually detected by the rennet test. Deficiency in bacterial growth factors may be one factor in causing slow starter. This phenomenon is also correlated with rapid reduction of methylene blue and high total count, but the real factor responsible is probably abnormal chemical composition.

VI. The incidence of mastitis increases steadily with age of the animal. Older cows do not give milk of higher “udder count” than younger cows if infection is absent. The increased count is entirely due to the increased incidence of mastitis. There appears to be a somewhat lower udder count in late lactation in infected cows, but a slightly higher one if mastitis is absent.

Finally we may emphasize the fact that for eradication purposes there are only two methods of diagnosis worth using—simple tests on the farm and carefully controlled bacteriological examination in the laboratory.